DEC RAID Array 3000 storage subsystem, DS-SWXRA-GA, DS-SWXRA-GC, DS-HSZ22-AA, DS-SWXRA-GD User Manual
...
RAID Array 3000 Storage
Subsystem
Hardware User’s Guide
EK–SMCPO–UG. A01
Digital Equipment Corporation
Maynard, Massachusetts
First Edition, January 1998
The disclosure of this information does not grant to the user a license under any patents,
pending patents, trademarks, or copyrights or other rights of Digital Equipment
Corporation, or of any third party.
This software is proprietary to and embodies the confidential technology of Digital
Equipment Corporation.
Possession, use or copying of the software described in this publication is authorized
only pursuant to a valid written license from Digital Equipment Corporation or an
authorized sublicensor.
Digital Equipment Corporation makes no representation that the use of its products in
the manner described in this publication will not infringe on existing or future patent
rights, nor do the descriptions contained in this publication imply the granting of
licenses to make, use, or sell equipment or software in accordance with the description.
The following are trademarks of Digital Equipment Corporation:
DEC, RAID Array 3000, StorageWorks, and the Digital Logo.
UNIX is a registered trademark in the United States and other countries, licensed
exclusively through X/Opin Company, Ltd..
All other trademarks and registered trademarks are the property of their respective
owners.
©
Digital Equipment Corporation 1997. All Rights Reserved
Printed in USA
Contents
Revision Record
About This Guide
..............................................................................................ix
...........................................................................................xi
1 Product Overview
1.1 Product Description...............................................................................1–1
1.2 Pedestal Features...................................................................................1–4
1.3 Pedestal Cabinet....................................................................................1–4
1.4 Pedestal Components ............................................................................1–6
1.4.1 StorageWorks Building Blocks (SBBs)..............................................1–6
1.4.2 RAID Array Controller......................................................................1–6
1.4.3
1.4.4
1.4.5
1.4.6
1.5 Specifications......................................................................................1–12
Pedestal Power Supplies ....................................................................1–7
Uninterruptable Power Supply...........................................................1–7
Environmental Monitor Unit (EMU) ..................................................1–8
UltraSCSI Bus ...................................................................................1–9
2 RAID Array Controller
2.1 Controller Overview.............................................................................. 2–1
2.2 Controller Features................................................................................2–2
2.3 Controller Reset and LED Indicators .....................................................2–6
2.4 Flexible RAID Set Configuration..........................................................2–7
2.5 Performance Enhancements ..................................................................2–8
2.5.1 Custom Components..........................................................................2–8
2.5.2 Efficient Write and Read Algorithms.................................................2–8
2.6 RAID Levels Supported......................................................................2–10
2.6.1 RAID 0............................................................................................ 2–11
2.6.2 RAID 1............................................................................................ 2–13
2.6.3 RAID 0+1........................................................................................2–13
2.6.4 RAID 4............................................................................................ 2–15
EK–SMCPO–UG. A01 iii
RAID Array 3000 Storage Subsystem
2 RAID Array Controller (continued)
2.6.5 RAID 5............................................................................................2–16
2.6.6 JBOD ..............................................................................................2–18
2.7 System Parameters.............................................................................. 2–18
2.8 Redundant Operation..........................................................................2–20
2.8.1 Initialization....................................................................................2–20
2.8.2 Message Passing.............................................................................. 2–20
2.8.3 Failover...........................................................................................2–20
2.9 Environmental ....................................................................................2–21
2.9.1 Backup Power Management............................................................2–21
2.9.2 Voltage Monitoring......................................................................... 2–22
2.9.3 Temperature Monitoring..................................................................2–24
3 Maintenance
3.1 Introduction..........................................................................................3–1
3.2 Pedestal Status and Power LEDs........................................................... 3–2
3.3 SBB Status LEDs.................................................................................. 3–2
3.4
3.5
3.6 EMU Error Conditions.......................................................................... 3–4
3.6.1 Replacing Components (FRU’s)........................................................3–5
3.6.2 Removing the Pedestal Door .............................................................3–5
3.6.3 Replacing an SBB .............................................................................3–6
3.6.4 Replacing a Power Supply................................................................. 3–7
3.6.5 Replacing the RAID Array Controller ...............................................3–8
3.6.6 Replacing the EMU Board.................................................................3–9
3.6.7 Replacing the UPS .......................................................................... 3–11
3.7 Differential/Wide UltraSCSI Bus........................................................3–11
3.7.1 Reconfiguring the SCSI Bus............................................................3–12
3.8 Replacing the Controller Memory Cache Modules..............................3–16
Controller LEDs ...................................................................................3–3
EMU Error Reporting ...........................................................................3–4
iv EK–SMCPO–UG. A01
4 Expansion Pedestal Option
4.1 Product Description...............................................................................4–1
4.2 Expansion Pedestal Cabinet...................................................................4–2
4.3 Expansion Pedestal Components...........................................................4–4
4.4 Reconfiguring Base Pedestal UltraSCSI Bus .........................................4–5
5 Second Controller Option
5.1 Introduction...........................................................................................5–1
5.2 Installation Procedure............................................................................ 5–2
Figures
1–1 RAID Array 3000 Pedestal Enclosure (Drives Optional) .......................1–2
1–2 Pedestal Front Panel Major Components...............................................1–5
1–3 Pedestal Rear Panel Power Supplies......................................................1–5
1–4 EMU Circuit Board Location................................................................1–8
1–5 Pedestal Rear Panel Components...........................................................1–9
1–6 UltraSCSI Bus Configuration Switch...................................................1–10
1–7 Slot Locations and SCSI ID Addresses................................................1–10
1–8 Slot Locations and SCSI ID Addresses................................................1–11
2–1 Bridging the Gap Between the Host and the Pedestal ............................2–2
2–2 Units Created from Storagesets, Partitions, and Disk Drives..................2–3
2–3 Controller Front Panel...........................................................................2–7
2–4 RAID 0 Write......................................................................................2–12
2–5 Diagram of a RAID 1 Write................................................................2–13
2–6 Diagram of a RAID 0+1 Write............................................................2–14
2–7 Diagram of a RAID 4 Write................................................................2–15
2–8 Diagram of a RAID 5 Write................................................................2–17
3–1 Pedestal Status LEDs ........................................................................... 3–2
3–2 Disk Drive Status LEDs ........................................................................3–3
3–3 Removing Pedestal Door.......................................................................3–5
3–4 Replacing an SBB.................................................................................3–7
3–5 Replacing a Power Supply.....................................................................3–8
3–6 Removing the Controller from the Pedestal...........................................3–9
3–7 Remove EMU Connector Nuts............................................................3–10
Contents
EK–SMCPO–UG. A01 v
RAID Array 3000 Storage Subsystem
Figures (continued)
3–8 Remove Screw and Panel.................................................................... 3–14
3–9 Location of SCSI Bus Configuration Switch.......................................3–15
3–10 Configuration Switch..........................................................................3–15
3–11 Remove Controller..............................................................................3–16
3–12 Release Locking Clips ........................................................................3–17
3–13 Remove Installed Memory Modules....................................................3–17
3–14 Install Replacement Modules.............................................................. 3–18
3–15 Pivot Module Down to Secure ............................................................3–19
4–1 Expansion Pedestal............................................................................... 4–1
4–2 Expansion Pedestal Slot Locations and ID Addresses............................4–3
4–3 Rear Panel Power Supplies ...................................................................4–4
4–4 EMU Circuit Board Location................................................................4–5
4–5 Remove Side Cover from Base Pedestal ...............................................4–7
4–6 Remove SCSI Bus Terminator.............................................................. 4–7
4–7 Disconnect SCSI Cable.........................................................................4–8
4–8 Connect SCSI Jumper...........................................................................4–8
4–9 Remove Connector Knockout Plate.......................................................4–9
4–10 Connect SCSI Cable ...........................................................................4–10
4–11 Set Configuration Switch to 7............................................................. 4–10
4–12 Configuration Switch..........................................................................4–11
4–13 Reconfigured SCSI Bus Addresses...................................................... 4–11
4–14 Connect SCSI Cable Between Pedestals.............................................. 4–12
4–15 Power Cable Connections...................................................................4–13
4–16 Transfer Drives from Base to Expansion Pedestal............................... 4–14
5–1 Second Controlle Slot Position..............................................................5–2
5–2 Insert Module into SIMM Connector ....................................................5–3
5–3 Pivot Module Down to Seat..................................................................5–4
5–4 Remove Controller from Top Slot.........................................................5–4
vi EK–SMCPO–UG. A01
Tables
1–1 RAID Array 3000 Part Numbers and Model Descriptions......................1–3
1–2 Pedestal Specifications........................................................................1–12
1–3 Pedestal Physical and Power Specifications.........................................1–13
2–1 Controller Specifications.......................................................................2–4
2–2 LED/Reset Switch Interface..................................................................2–6
2–3 RAID Levels Supported......................................................................2–10
2–4 Pedestal RAID Set Restrictions...........................................................2–10
2–5 RAID 0+1 Example.............................................................................2–14
2–6 System Parameters..............................................................................2–19
2–7 Backup Power Management................................................................ 2–21
2–8 Acceptable System Voltage Levels .....................................................2–22
2–9 Acceptable Termination Voltage Levels..............................................2–23
2–10 Acceptable 12 Volt Levels ..................................................................2–23
2–11 Acceptable External Temperature Voltage Levels............................... 2–24
2–12 Acceptable Board Temperature Voltage Levels................................... 2–24
3–1 Disk Drive SBB Status LEDs................................................................3–3
3–2 SCSI Bus Length and External Cables.................................................3–11
3–3 Assigned Slot Device Addresses in the Pedestal..................................3–11
Contents
EK–SMCPO–UG. A01 vii
Revision Record
This Revision Record provides a concise publication history of this guide. It lists the
revision levels, release dates, and reasons for the revisions.
The following revision history lists all revisions of this publication and their
effective dates. The publication part number is included in the Revision Level
column, with the last entry denoting the latest revision. This publication supports
the StorageWorks RAID Array 3000 Storage Subsystem.
Revision Level Date Summary of Changes
EK–SMCPO–UG. A01 January 1998 Original release.
EK–SMCPO–UG. A01 ix
About This Guide
This section identifies the audience of this guide and describes the contents (chapter-bychapter) and structure. I n addition, this section includes a list of associated documents
and the conventions used in this guide.
Intended Audience
This guide is intended for installers and operators of the RAID Array 3000
storage subsystem. Installing the subsystem requires a general understanding of
basic SCSI terminology and SBus product installation procedures.
Document Structure
This guide contains the following chapters:
Chapter 1: Product Overview
Product Overview
the single pedestal RAID Array 3000 storage subsystem. It includes the major
features, a brief description of the major components, and the specifications for
the pedestal.
Chapter 2: RAID Array Controller
RAID Array Controller
controller in the RAID Array 3000. It also lists the RAID levels supported by the
subsystem and a brief description of each level. Redundant operation and
environmental considerations (i.e. backup power management, voltages, and
temperature) are covered at the end of the chapter.
provides an overview and a physical hardware description of
describes the major features and characteristics of the
Chapter 3: Maintenance
Maintenance
pedestal as a troubleshooting aid in case of a problem. The chapter also describes
how to replace a Field Replaceable Unit (FRU) in the pedestal, how to reconfigure the SCSI bus, and how to add or replace a memory SIMM module in
the RAID controller.
EK–SMCPO–UG. A01
describes how to interpret the status and power LEDs on the
xi
RAID Array 3000 Pedestal Enclosure
Chapter 4: Expansion Pedestal Option
Expansion Pedestal Option
the pedestal expansion kit option. It also explains how to reconfigure the SCSI
bus in the base pedestal to accommodate the added storage capability of the
expansion pedestal.
describes the major features and characteristics of
Chapter 5: Second Controller Option
Second Controller Option
pedestal for redundancy.
describes how to install a second controller in the
Associated Documents
In addition to this guide, the following documentation is useful to the reader:
Table 1 Associated Documents
Document Title Order Number
StorageWorks RAID Array 3000 Configuration and
Maintenance Guide
Installation Instructions for the RAID Array 3000 Pedestal
Expansion Kit Option Product Note
Installation Instructions for the RAID Array Replacement
Controller – SWXRC-03
EK–SMCS2–UG
EK–SMCPM–PN
EK–SMCPL–PN
Conventions
This guide uses the following documentation conventions:
Table 2 Style Conventions
Style Meaning
plain monospace type Text
boldface type
italic type For emphasis, manual titles, chapter
xii
For the first instance of terms being
defined in text, or both.
summaries, keyboard key names
.
EK–SMCPO–UG. A01
About This Guide
Conventions (
continued
)
Table 3 Nomenclature Convention
RAID Advisory Board Description RAID Array 3000 Usage
RAID 0 STRIPset
RAID 1 MIRRORset
RAID 0+1 STRIPED MIRRORset
RAID 4 STRIPED with a Fixed parity drive
RAID 5 STRIPED with a Floating parity drive
Support and Services
Who to contact in the Americas
Information and Product Questions:
Installation Support:
DIGITAL Multivendor Customer Service (MCS):
Installation
Warranty
Remedial
Customer Support Center (CSC)
Local Sales Office / StorageWorks Hotline
1-800-786-7967
Contact the DIGITAL Distributor where the
Storage Solution was Purchased / Local Digital
Sales Office.
Contact the DIGITAL Customer Support Center
(CSC).
Contact the DIGITAL Customer Support Center
(CSC) for warranty service after solution is
installed and operating.
Contact the DIGITAL Customer Support Center
(CSC)
A Service Contract is recommended when
Note:
the equipment is out of warranty. Contact the
local DIGITAL Sales Office.
1 800-354-9000
EK–SMCPO–UG. A01
xiii
RAID Array 3000 Pedestal Enclosure
Who to contact in Europe
Information and Product Questions: Contact the DIGITAL Distributor or reseller
Installation Support and Installation: Contact the DIGITAL Distributor or reseller from
whom the Storage Solution was purchased.
For Warranty Service See the Warranty Card packaged with the
product.
For Remedial Service Contact the DIGITAL Distributor or reseller from
whom the Storage Solution was purchased.
Note: A Service Contract is recommended
when the equipment is out of warranty.
Who to contact in Asia Pacific
For all services, contact the DIGITAL Distributor or reseller from whom the
equipment was purchased
xiv
EK–SMCPO–UG. A01
1
Product Overview
This chapter provides an overall description of the RAID Array 3000 storage subsystem
and its components. A list of technical and environmental specifications is located at the
end of the chapter.
NOTE
This guide is the Hardware User’s Guide. For c onfiguration information, refer to the
RAID Array 3000 for W indows NT – Int el Installation Guide
sole (SWCC) 2.0 User’s Guide.
and the
StorageW orks Command Con-
1.1 Product Description
The RAID Array 3000 storage subsystem (Figure 1–1) is a desk-side subsystem
offering the basic components required to create a user-designed storage array
with two 16-bit, differential UltraSCSI bus host interfaces. The pedestal can accommodate up to seven 3½-inch storage devices. The devices, referred to as
StorageWorks Building Blocks or SBBs, are disk drives from the StorageWorks
family of storage devices. The release note that accompanies the subsystem lists
the software solutions and disk drives that are supported. A battery backup subsystem is included as part of the pedestal enclosure in the form of a free-standing
UPS (Uninterruptable Power Supply). In case of a power failure, the UPS provides a temporary backup for cache while the subsystem flushes to disks.
The RAID Array 3000 offering also includes option kits designed to increase the
storage capacity and enhance the performance of the subsystem. The first is an
expansion pedestal (second enclosure) designed to increase the storage capacity
of the subsystem to a maximum of 14 drives. The pedestal expansion kit option
is described in detailed in Chapter 4 of this guide.
The second option allows the addition of a second RAID controller to the subsystem for redundancy. The second controller operates in conjunction with the installed controller to protect data during a malfunction. Chapter 5 describes how
to install the redundant controller option and how to reconfigure the subsystem to
accommodate it.
Getting Started
EK–SMCPO–UG. A01 1–1
RAID Array 3000 Storage Subsystem
The RAID Array 3000 pedestal enclosure and its associated options are listed
and described in Table 1–1. Figure 1–1 shows the pedestal with a full complement of drives (optional) for completeness.
Figure 1–1 RAID Array 3000 Pedestal Enclosure (Drives Optional)
PEDESTAL
The RAID 3000 pedestal is equipped with a dual-channel RAID controller which
supports all of the UltraSCSI bus features. It also contains an Environmental
Monitor Unit (EMU) board for environmental monitoring and error detection.
1–2 EK–SMCPO–UG. A01
UPS
3000-01A
Chapter 1. Product Overview
y
y
Table 1–1 Pedestal RAID Array 3000 Part Numbers and Model Descriptions
DIGITAL Part No. Item Description
DS-SWXRA-GA
DS-SWXRA-GC
DS-HSZ22-AA Second HSZ22 controller with 16 MB of cache and two 16 MB
DS-SWXRA-GD Expansion Pedestal (120/240 V) with slot space for seven addi-
DS-SWXRA-GR Single 204-watt power supply for RA3000, 120/230 V for on-site
RA3000 pedestal subsystem with one controller, 120 V.
Includes:
HSZ22 two-channel controller with 16 MB cache, Environmental
Monitor Unit (EMU), two 204 watt power supplies with fans, five
meter host SCSI cable (BN37A), BN38E-OB adapter, one 120volt Un-interruptable power suppl
power cords. Disks are not included.
Requires
disks.
Options
pedestal, and cache memory upgrade.
RA3000 pedestal subsystem with one controller, 230 V.
Includes:
HSZ22 two-channel controller with 16 MB cache, Environmental
Monitor Unit (EMU), two 204 watt power supplies with fans, five
meter host SCSI cable (BN37A), BN38E-OB adapter, one 230volt Un-interruptable power suppl
power cords. Disks are not included.
Requires
disks.
Options
pedestal, and cache memory upgrade.
SIMMs for mirror cache, one BN37A-05 five-meter UltraSCSI
cable, and one BN38E-OB adapter.
tional UltraSCSI disk drives.
spare.
Seven-slot pedestal for wide UltraSCSI SBBs, one
(UPS), and North American
: Solutions Software Kit for platform, host adapter, and
: Second HSZ22 controller, seven-disk SBB expansion
Seven-slot pedestal for wide UltraSCSI SBBs, one
(UPS), and North American
: Solutions Software Kit for platform, host adapter, and
: Second HSZ22 controller, seven-disk SBB expansion
EK–SMCPO–UG. A01 1–3
RAID Array 3000 Storage Subsystem
1.2 Pedestal Features
The major features of the pedestal are:
Two differential 16-bit UltraSCSI host buses
•
Seven 3½-inch disk drive SBB slots
•
One dual-channel RAID array controller
•
Second controller option for redundancy
•
Expansion pedestal option allowing up to fourteen SBB slots in a dual-
•
pedestal subsystem configuration
Memory cache expansion option for the controller
•
Redundant power provided by two fan-cooled universal ac input power sup-
•
plies (50/60 Hz, 100 to 240 V ac)
Cache backup provided by an external Uninterruptable Power Supply (UPS)
•
Environmental monitor unit (EMU) for error detection
•
The ability to hot swap SBBs without powering down the system
•
1.3 Pedestal Cabinet
The pedestal cabinet is a modular free-standing storage enclosure that is completely self contained. It has two fan-cooled power supplies, an internal EMU
circuit board, and a RAID array controller with front panel display and control.
Figure 1–2 shows the major components in the pedestal enclosure.
Figure 1–3 identifies the items on the rear panel power supplies. The character-
istics of the pedestal cabinet are outlined below.
The disk drive storage capacity is seven 3½-inch disk drive SBBs
•
The subsystem slots are numbered 0 through 6 from top to bottom
•
There are two 68-pin high density female SCSI connectors on the rear panel
•
which interconnect the host system to the RAID controller in the pedestal
The rear panel also contains an alarm switch, a UPS monitor connector, an
•
external fault condition connector, and a serial port connector (for controller
configuration)
The pedestal is equipped with an internal configuration switch which sets the
•
SCSI ID addresses of the controller and the storage devices
1–4 EK–SMCPO–UG. A01
Chapter 1. Product Overview
Figure 1–2 Pedestal Front Panel Major Components
Figure 1–3 Pedestal Rear Panel Power Supplies
EK–SMCPO–UG. A01 1–5
RAID Array 3000 Storage Subsystem
1.4 Pedestal Components
The major components in the pedestal subsystem include:
Dual-channel RAID array controller
•
Two 16-bit single-ended split SCSI buses
•
Environmental Monitor Unit (EMU)
•
Two universal 50/60 Hz, 120 or 240 Vac power supplies
•
Separate free-standing backup power supply (UPS)
•
1.4.1 StorageWorks Building Blocks (SBBs)
The pedestal has seven 3½-inch disk drive SBB slots. The number of drives that
make up each configuration of the array is left to the end user with a maximum
of seven storage devices.
1.4.2 RAID Array Controller
The controller contains two Wide/UltraSCSI/differential host channels and two
Wide/UltraSCSI/Single-Ended disk channels. In dual-controller configurations,
the controllers support fully automatic and smooth controller failover.
The controller supports one or two standard 72-pin cache SIMMs of up to
64 MB. In a redundant controller setup, both controllers must have identical
cache configurations and the total usable cache (per controller) will be half the
amount installed. Thus, in a single controller setup the maximum usable cache is
128 MB while a redundant setup has a maximum usable cache of 64 MB (per
controller).
The RAID Array controller contains the following features:
Single PCB form factor for inclusion in the enclosure
•
Support for dual hot-swap controller operation
•
Dual differential Ultra-Wide SCSI host channels
•
Dual single-ended Ultra-Wide SCSI disk channels
•
RAID level 0, 1, 0+1, 4, 5, and JBOD support
•
EMU support
•
Cluster support for Windows NT
•
32 Logical Units (LUNs) per host channel (some operating systems may be
•
limited to 8)
Support for Hot and Warm spare disks
•
UPS backed write caching
•
1–6 EK–SMCPO–UG. A01
Chapter 1. Product Overview
Per LUN write cache/write back selection
•
Configuration/Maintenance via RS-232 or host SCSI channel using SWCC
•
(StorageWorks Command Console)
Update of firmware via host channel
•
1.4.3 Pedestal Power Supplies
The pedestal has two interchangeable, air-cooled, AC power supply modules located at the rear of the unit. The power supplies provide redundant power if one
of the units should malfunction. Each supply provides +5 and +12 Vdc to power
the RAID controller, EMU, and the storage SBBs in the pedestal. In addition,
each unit contains a high-speed fan for pedestal cooling. The upper power supply
is designated as "A" and the lower as "B". The unit contains an ac power receptacle, a power status LED, a fan, and a latching slider switch to secure the supply
in the pedestal.
1.4.4 Uninterruptable Power Supply (UPS)
The UPS is separate and free-standing unit designed to protect the pedestal from
problems associated with poor quality AC power or a complete loss of AC
power. The UPS is connected between the AC outlet and the line input of the
pedestal power supply “B” to provide battery backup power.
The major features of the UPS include Cell Saver Technology (doubles battery
life and speeds recharge time), hot-swap batteries, and network surge protection.
The front panel display has user controls (LEDs and control buttons) and the rear
panel contains a COMM port, which provides UPS status to the EMU in the pedestal. The rear panel also contains the network surge protector, a reset button,
and four power receptacles. An audible alarm is activated when input power
fails, as a Low Battery Warning, or whenever the UPS is in need of servicing.
The UPS automatically recharges its battery when power is returned following a
power failure. Recharge time is four to six hours depending on the energy requirements of your load and the length of the power outage.
Its own installation, operation, and service manual support the UPS. The manual
describes the UPS in detail and is part of the documentation set enclosed with
your subsystem.
EK–SMCPO–UG. A01 1–7
RAID Array 3000 Storage Subsystem
300
0-3
1.4.5 Environmental Monitor Unit (EMU)
The EMU is an internal circuit board that monitors the operation of the pedestal.
The EMU monitors power supply voltages, fans, temperatures which are reported
to the user, and controls (turns on and off) the audible alarm and status LED on
the front panel. The EMU also reports the subsystem status to the controller that
reports to the host, and has the capability of exchanging signals with auxiliary
devices and controllers.
The EMU is located internally in the top rear of the pedestal as shown in Figure
1–4). It is connected to the SCSI bus and powered by an internal cable. The following external components on the rear panel of the pedestal are part of the
EMU (see Figure 1–5):
An alarm switch (S1) that enables (up) or disables (down) the audible alarm
•
A power monitor connector (UPS) allows the EMU to monitor and report the
•
status of a battery backup power supply
An External Fault Condition connector allows the EMU to monitor the status
•
of a user-selected device
Figure 1–4 EMU Circuit Board Location
EMU
1–8 EK–SMCPO–UG. A01
7
Figure 1–5 Pedestal Rear Panel Components
300
0-23
)
)
Chapter 1. Product Overview
UPS
Host #0
External
Fault
Host #1
Contro ller (To p
Controller (Bottom
A
1.4.6 UltraSCSI Buses
The pedestal contains two, 16-bit, single-ended, wide UltraSCSI buses (factoryconfigured as a split bus) that connects the controller(s) to the disk drives.
The split-bus arrangement is divided into two bus paths designated
and
device port 1
upper four devices in the pedestal (IDs 8 through 11) and port 1 connects the
controller to lower three devices (IDs 8 through 10). The device addresses on the
bus are set at the factory by an internal configuration switch (see Figure 1–7).
EK–SMCPO–UG. A01 1–9
as shown in Figure 1–6. Port 0 connects the controller to the
device port 0
RAID Array 3000 Storage Subsystem
Figure 1–6 UltraSCSI Bus Port and Default SCSI ID Assignments
ID 8
ID 9
ID 1 0
ID11
ID 8
Device
Port 0
Top Controller
ID =7
B o ttom Con troller (Op t io na l)
ID =6
ID 9
Device
Port 1
ID 1 0
ID = 7
ID = 6
300 0- 51
When set to a specific position, the switch controls the addresses of each SBB
slot. Figure 1–8 identifies the pedestal slot locations and their corresponding
SCSI ID addresses for each device port.
The subsystem can be reconfigured for “through-bus” operation by modifying the
bus and resetting the internal configuration switch. The information needed to
reconfigure the bus from “split-bus” to a “through-bus” configuration is described in Chapter 4 of this guide (
Expansion Pedestal Option
).
1–10 EK–SMCPO–UG. A01
Figure 1–7 UltraSCSI Bus Configuration Switch
300
0-3
8
Figure 1–8 Slot Locations and SCSI ID Addresses
Chapter 1. Product Overview
6
EK–SMCPO–UG. A01 1–11
SLOT 0
SLOT 1
SLOT 2
SLOT 3
SLOT 4
SLOT 5
SLOT 6
Controller
ID = 8
ID = 10
Device
Port 0
11
8
Device
Port 1
-
RAID Array 3000 Storage Subsystem
1.5 Specifications
Table 1–2 Pedestal Technical Specifications
Technical Specifications
Item Description
Cabinet Pedestal with seven (7) disk SBB slots
Controller HSZ22
Controller cache 16 MB standard
Backup for cache Standard un-interruptable power supply (UPS)
Mirrored write-back cache Yes
Device channels per controller 2
Maximum disks per device
port
Dual active controllers Yes, order HSZ22-Aa and second solutions
Host interface UltraSCSI wide differential
Drive interface UltraSCSI wide single-ended
RAID levels supported 0, 1, 0+1, 4, 5
Non-RAID disk support Yes (JBOD)
Sustained I/O rate 4,400 I/O’s per second per controller pair
RAID 5 sustained transfer rate 28 MB per second per controller pair
Maximum transfer rate 40 MB per second per controller pair
Redundant fans Yes
Redundant power supplies Yes
Global disk spares Yes
Environmental Monitoring Unit Yes, monitors power and temperature
Setup/control lines One serial
Serviceability Hot-swap components
RAID manager GUI support StorageWorks Command Console (SWCC)
Regulatory approvals UL, CSA, TUV, FCC, CE MARK, C TICK,
Expansion pedestal with an additional 7 slots
Upgrades to 128 MB for a two controller pair
14
software kit
2.0 available for all platforms. Netware requires serial line connection
BCIQ, VCCI
1–12 EK–SMCPO–UG. A01
Chapter 1. Product Overview
y
Table 1–3 Pedestal Physical and Power Specifications
Physical Specifications
Item Dimension
Height 564 mm
Width 254 mm
Depth 494 mm
Rear Clearance (air exhaust) 305 mm
Front Clearance (door opening) 305 mm
Weight (no devices) 19.5 kg
Power Specifications
Item Rating
Input power 110-240 Vac, 50/60 Hz, single
phase, 12A/6A
Heat dissipation 3070 BTUs/hr.
Temperature (optimal, minimum required) 18 to 24° C, 10 to 40° C
Altitude Up to 2,400 m
Air quality Not to exceed 500,000 parti-
cles/ft3 for air at a size of 0.5
micron or larger
Total power per power suppl
(Total of +12 Vdc and +5 Vdc outputs)
Nominal output voltages +5 Vdc @ 15 amps, maximum
Device startup time 4-second internal, minimum
204 Watts, maximum
+12 Vdc @ 12 amps, maximum
EK–SMCPO–UG. A01 1–13
2
RAID Array Controller
This chapter describes the major featur es and characteristics of the RAID array controller in the RAID Array 3000 subsystem. The num ber of devices suppor ted by t he cont roller may be limited by the enclosure.
2.1 Controller Overview
The RAID Array controller provides high performance, high-availability access
to SCSI disk array subsystems along a UltraSCSI/Wide SCSI bus. With a modular hardware design and an intuitive configuration utility, the controller is designed to meet a wide range of storage needs.
The controller consists of a single 5 1/2” x 8” PCB mounted in a sheet metal
subassembly. The package consists of the controller PCB, a 300-pin connector,
mechanical insertion assists, and an LED/reset switch interface. All signals to the
controller are routed through the backplane connector.
The unit is configured with two Ultra Wide, differential SCSI host channels capable of transferring data to and from the host at rates up to 40 MB/s. The host
SCSI IDs are configurable via the Host Parameters and can support 32 deep
tagged queuing. The controller is also configured with two Ultra Wide, singleended SCSI disk channels capable of transferring data to and from the disk drives
at rates up to 40 MB/s. Each channel can support up to 15 drives (14 in redundant controller subsystems).
The controller has two SIMM connectors for up to 128 MB of cache memory.
The SIMM connectors form a mirrored pair when the controllers are configured
in a redundant controller configuration; otherwise they are fully accessible by the
controller. In a redundant controller setup, both controllers must have identical
cache configurations and the total usable cache (per controller) will be half the
amount installed. Thus, in a single controller setup the maximum usable cache is
128 MB while a redundant setup has a maximum usable cache of 64 MB (per
controller).
EK–SMCPO–UG. A01 2–1
RAID Array 3000 Pedestal Enclosure
There are two configurations for redundant pairs of controllers:
Failover
mode and
Active/Passive Failover
mode. In Active/Active Failover,
Active/Active
each controller in the redundant pair has one active SCSI host port and one passive SCSI host port. Redundancy Groups (Virtual LUNS) can be mapped only to
one active host port and are not accessible from the passive port or the other
controller (i.e. partitioned model).
In Active/Passive Failover, one controller in the redundant pair has both SCSI
host ports active and and the other controller is in a standby passive mode. Redundancy Groups (Virtual LUNS) can be mapped to either SCSI host port or to
both as in the single controller model.
In both cases a single controller failure will not affect the subsystem because the
surviving controller will take over.
2.2 Controller Features
The controller is the intelligent bridge between the host and the devices in the
pedestal. From the host’s perspective, the controller is simply another SCSI device connected to one of its I/O buses. Consequently, the host sends its I/O requests to the controller just as it would to any other SCSI device. Figure 2–1
shows the role of the controller between the host and the pedestal.
Figure 2–1 Bridging the Gap Between the Host and the Pedestal
2–2 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
From the pedestal’s perspective, the controller receives the I/O requests from the
host and directs them to the devices in the pedestal. Since the controller processes all the I/O requests, it eliminates the host-based processing that is typically
associated with reading and writing data to multiple storage devices.
The controller does much more than simply manage I/O requests: it provides the
ability to combine several ordinary disk drives into a single, high-performance
storage unit called a storageset. Storagesets are implementations of RAID technology, also known as a “Redundant Array of Independent Disks”. Every storageset shares one important feature: whether it uses two disk drives or 14, each
storageset looks like a single storage unit to the host.
You create storage units by combining disk drives into storagesets such as stripesets, RAIDsets, and mirrorsets, or by presenting them to the host as single-disk
units (see Figure 2–2).
Figure 2–2 Units Created from Storagesets, Partitions, and Disk Drives
EK–SMCPO–UG. A01 2–3
RAID Array 3000 Pedestal Enclosure
y
Stripesets (RAID 0) combine disk drives in serial to increase transfer or re-
•
quests rates
Mirrorsets (RAID 1) combine disk drives in parallel to provide a highly reli-
•
able storage unit
RAID 4 provides striping with a fixed parity drive
•
RAIDsets (RAID 5) combine disk drives in serial - just like stripesets - but
•
also store parity data to ensure high reliability
Stripe mirrorsets (RAID 0 + 1) combine mirrorsets in serial to provide the
•
highest throughput and availability of any storage unit
Table 2–1 summarizes the features of the RAID 3000 controller.
Table 2–1 Controller Specifications
Item
Environmental Monitoring High Availability Fault Bus support via EMU
Processor 40 MHz, 32 bit LR33310 RISC CPU
System Bus Interface Two, WIDE, Differential UltraSCSI channels
Configuration Two RS232 serial ports
RAID Levels Supported 0, 1, 0+1, 4, 5
Drive Channels Supported Two, WIDE, UltraSCSI single-ended channels
Number of Logical Drives (LUN’s) Up to 30 RAID sets, and up to 16 redundanc
SCSI Channels Two, UltraSCSI, 16-bit, single-ended
Metadata 20 blocks/disk (10240 bytes)
Largest Allowable Disk/RAID set/LUN Two, 32 blocks (approximately 2.2 petabytes)
Non-RAID Device Support Disk Drives (JBOD)
Drives Supported StorageWorks 2, 4, and 9m GB SCSI and Ul-
Drive Reconstruct Automatic with hot or warm spares
Disk Hot Swap Yes
Disk Hot Spare (spinning) Yes, global hot spare
Specifications
groups (LUNs) per RAID set
traSCSI drives.
2–4 EK–SMCPO–UG. A01
Table 2–1 Controller Specifications (Continued)
y
Item Specifications
Disk Warm Spare (not spinning) Yes, global warm spare
Redundant Power Supplies Yes
Redundant Controllers Yes
Controller Failover Yes, automatic
Controller Hot Spare Yes (Active-Passive mode)
Controller Hot Swap Yes
Cluster Support Yes, Single (SCSI) bus cluster
Maximum number of units presented to
host
Maximum host port transfer speed 20 MHz
Command Queuing Yes, 64 commands (host and disk SCSI chan-
Heterogeneous Multi-Host Support Yes
Mixed Drive Types Yes
Configurable Reconstruct Time Yes
Stripe Size (chunk size) Variable
Write through Cache Yes, User Selectable (default)
Write Back Cache Yes, User Selectable (optional)
Write on Top Yes
Write Gathering Yes
Battery Backup for Cache Yes, Pedestal Uninterruptable Power Suppl
Boot Capability Bootable from RAID set (System dependent)
Number of Controllers /System Two
Cache Support Up to 256 MB (using two 128 MB industry-
FCC Rating Class B
Temperature 5ºC to 50ºC operating, -40ºC to +60ºC non-
Relative Humidity 10% to 95% non-condensing (operating), 5% to
Physical Size 8.55” deep, 5.03” wide, 1.6” high
Power Requirements 5 VDC @ 3 A, peak; - 12 VDC @ 1 A, peak
64
nels)
(UPS)
standard, 72-pin, 36-bit, 60 ns SIMMS)
Environmental
operating
90%, non-condensing (non-operating)
Chapter 2. RAID Array Controller
EK–SMCPO–UG. A01 2–5
RAID Array 3000 Pedestal Enclosure
2.3 Controller Reset and LED Indicators
Figure 2–3 illustrates the front panel of the controller. All LEDs are numbered
from left to right. The reset button (LED 0) flashes green about once every second (heartbeat) to indicate that the controller is operating normally. LEDs 1
through 4-display host and disk channel activity (amber). LED 5 (normally off)
comes on red during a controller failure. The LED/Reset switch interface is defined in Table 2–2.
Table 2–2 LED/Reset Switch Interface
LED # Name
0 Heart Beat Controller Reset Switch (green)
1 Host Channel 0 Activity LED (amber)
2 Host Channel 1 Activity LED (amber)
3 Disk Channel 0 Activity LED (amber)
4 Disk Channel 1 Activity LED (amber)
5 Fault LED (red)
2–6 EK–SMCPO–UG. A01
Figure 2–3 Controller Front Panel
Chapter 2. RAID Array Controller
2.4 Flexible RAID Set Configuration
In addition to its flexible hardware design, the controller’s firmware offers the
user the flexibility to configure RAID sets in several different ways:
RAID sets can comprise drives from any drive channel and SCSI ID.
•
A RAID set can contain all the drives connected to the controller, a single
•
drive, or any number of drives in between.
The controller supports RAID Levels 0, 1, 0+1, 4, and 5. It also supports
•
JBOD (Just a Bunch of Drives), allowing you to connect standalone disk
drives (such as a system disk) to the controller without making them members of a RAID set.
EK–SMCPO–UG. A01 2–7
RAID Array 3000 Pedestal Enclosure
Each RAID set can be partitioned into smaller redundancy groups.
•
The controller’s host LUN Mapping feature makes it possible to map RAID
•
sets differently to each host port. You make the same redundancy group appear on different LUNs to different hosts, or make a redundancy group visible to one host but not to another.
Any drive may be designated as a hot or warm spare. Spares are global,
•
meaning that in the event of a drive failure, the controller will search for the
first available spare on any channel or SCSI ID and automatically begin rebuilding the failed drive’s data.
2.5 Performance Enhancements
The controller employs a number of techniques to achieve as much performance
as possible from its design.
2.5.1 Custom Components
To increase performance and reliability, the controller’s core functions have been
encapsulated in four custom ASIC (Application Specific Integrated Circuits)
components as follows:
XOR ASIC:
levels 4 and 5.
DMA ASIC:
CPU Interface ASIC:
processing unit.
Memory Controller ASIC:
movement on the internal bus at a maximum burst rate of 80 MB/second and a
maximum sustainable rate of 60 MB/second.
Used in the Exclusive -Or parity calculations employed by RAID
Controls the data path hardware for the various I/O ports
Supports the controller’s MIPS R3000 RISC central
Controls the memory system and supports data
2.5.2 Efficient Write and Read Algorithms
Standard RAID write operations that involve parity, such as those in RAID levels
4 and 5, require multiple, time-consuming steps:
1. Read data from the parity drive.
2. Read existing data from the target data drives.
3. Exclusive-Or the old parity, old data, and new data to generate new parity
data.
4. Write the new parity data to the parity drive.
5. Write the new data to the target data drives.
2–8 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
The controller uses several techniques to streamline write operations and significantly improve performance. All the techniques use the controller’s on-board
cache, which can contain up to 64 MB of memory in the form of standard 72-pin,
60-nanosecond SIMMs.
NOTE
The controller will not operate wit hout at leas t one
4 MB SIMM installed in its cache. Nor will it operate without either a bac kup or an un- interr uptable
power supply connec ted to the c ontroll er. Wit hout
a backup, data stored in the cache, but not yet
written to the disk drives, would be lost in the
event of a power interruption.
2.5.2.1 Write-Back Caching
When the host sends data to be written to a redundancy group the controller
stores the data in its cache and immediately reports to the host it has completed
the write. The controller eventually writes the data to the disk drives when the
write can be done most efficiently, or when the controller must flush the cache to
make room for other data or to prepare for a shutdown.
Write-back caching makes the host more responsive to the user, since the host
does not have to wait for a lengthy RAID write before proceeding to another
task.
2.5.2.2 Write Gathering
The controller will attempt to consolidate multiple writes destined for contiguous
blocks and then write the entire data block in one operation. The controller stores
the data in cache until it performs the write. Ideally, the controller will wait until
it has gathered enough data to fill an entire stripe. This enables the controller to
avoid reading form the parity and data drives before making the write. All it has
to do is calculate parity from the data it already has in its cache, then write the
data and parity to the drives. Even if the controller cannot accumulate enough
data to fill a stripe, the consolidation of small writes can reduce the number of
read/write operations that must take place.
2.5.2.3 Write On Top
If the host commands that data be written to disk, and data for that address is
pending in the controller’s cache, the controller writes the new data on top of the
old in the cache. Only the new data is eventually written to the disk drives.
EK–SMCPO–UG. A01 2–9
RAID Array 3000 Pedestal Enclosure
2.6 RAID Levels Supported
The RAID Array 3000 controller supports the following RAID levels:
Table 2–3 RAID Levels Supported
RAID Level Description
0 Striping without parity
1 Mirroring
0+1 Striping and mirroring
4 Striping with fixed parity drive
5 Striping with floating parity drive
JBOD “Just a Bunch of Drives”
NOTE
The controller stripes data in multi-block chunk
sizes . Also, the cont roller does not support RAID
level 3 or 0 with a one-block chunk size.
There are some restrictions you must adhere to when creating a RAID set on the
RAID 3000 pedestal. The minimum and maximum number of drives required to
support each RAID level is listed in Table 2–4.
Table 2–4 Pedestal RAID Set Restrictions
RAID Level Min. No.1 of
Drives
JBOD 1 1
02 56
12 56
0+1 4 16
43 56
53 56
1
Must be even number.
2–10 EK–SMCPO–UG. A01
Max. No.1 of
Drives
Chapter 2. RAID Array Controller
2.6.1 RAID 0
RAID 0 breaks up data into smaller chunks and writes each chunk to a different
drive in the array. The size of each chunk is determined by the controller’s chunk
size parameter, which you set in the course of creating a RAID set.
The advantage of RAID 0 is its high bandwidth. By breaking up a large block of
data into smaller chunks, the controller can use multiple drive channels to write
the chunks to the disk drives. Furthermore, RAID 0 involves no parity calculations to complicate the write operation. Likewise, a RAID 0 read operation employs multiple drives to assemble a single, large data block. This makes RAID 0
ideal for applications such as graphics, video, and imaging that involve the writing and reading of large, sequential blocks. Figure 2–4 shows a diagram of a
RAID 0 write.
CAUTION
The lack of parity means t hat a RAID 0- disk arr ay
offers absolutel y no redundancy and thus cannot
recover from a drive failure.
EK–SMCPO–UG. A01 2–11
RAID Array 3000 Pedestal Enclosure
Figure 2–4 RAID 0 Write
2–12 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
2.6.2 RAID 1
RAID 1 (also known as mirroring or shadowing) takes data sent by the host and
duplicates it on all the drives in an array. The result is a high degree of data
availability, since you can lose all but one drive in the array and still have full
access to your data. This comes at a price: a RAID 1 array requires multiple
drives to achieve the storage capacity of a single drive. Figure 2–5 illustrates a
RAID 1 write.
Figure 2–5 Diagram of a RAID 1 Write
A RAID 1 array will show up on the monitor as “degraded” when at least one
drive fails, even if two or more members of the redundancy group remain in
good working order. As long as at least two working drives remain in the array,
you may continue to run the array in degraded mode without putting data in
jeopardy.
2.6.3 RAID 0+1
RAID 0+1 combines RAID 0 (striping) with RAID 1 (mirroring). In RAID 0+1
write, the controller breaks up the data block from the host into smaller chunks,
then writes the chunks to half the drives in the array, while writing duplicate
chunks to the remaining drives.
EK–SMCPO–UG. A01 2–13
RAID Array 3000 Pedestal Enclosure
Figure 2–6 Diagram of RAID 0+1 Write
In the event of a drive failure, a RAID 0+1 array will enter degraded mode and
continue to operate by substituting the failed drive with its mirror.
When the controller creates a RAID 0+1 set, it first sorts the drives by channel
number and SCSI ID. Then it stripes the data across every other drive and forms
a mirrored pair with the first two drives, another mirrored pair with the second
two drives, and so on. Table 2–5 describes how the controller uses the drives in a
RAID 0+1 set.
Table 2–5 RAID 0+1 Example
Drives Selected Function
Channel 1, ID 0 First member of stripe set.
Channel 1, ID 1 Mirror of channel 1, ID 0
Channel 1, ID 2 Second member of stripe set
Channel 2, ID 0 Mirror of channel 1, ID 2
Channel 2, ID 1 Third member of stripe set
Channel 2, ID 2 Mirror of channel 2, ID 1
2–14 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
2.6.4 RAID 4
RAID 4 breaks up host data into chunks, calculates parity by performing an exclusive-or on the chunks, and then writes the chunks to all but one drive in the
array and the parity data to the last drive. When the host requests data from the
disk drives, the controller retrieves the chunks containing the addressed data, reconstitutes the data from the chunks, and passes the data to the host.
Figure 2–7 Diagram of a RAID 4 Write
EK–SMCPO–UG. A01 2–15
RAID Array 3000 Pedestal Enclosure
In the event of a single drive failure, a RAID 4 array will continue to operate in
degraded mode. If the failed drive is a data drive, writes will continue as normal,
except no data will be written to the failed drive. Reads will reconstruct the data
on the failed drive by performing an exclusive-or operation on the remaining
data in the stripe and the parity for that stripe. If the failed drive is a parity drive,
writes will occur as normal except no parity will be written. Reads will simply
retrieve data from the data disks. There will be no deterioration in controller performance while a RAID set is in degraded mode.
In general, RAID 4 is best suited for applications such as graphics, imaging, or
video that call for reading and writing large, sequential blocks of data. However,
you may find that RAID 4 is preferable to RAID 5 even for applications characterized by many small I/O operations, such as transaction processing. This is due
to the controller’s intelligent caching, which efficiently handles small I/O reads
and writes, and to the relatively less complex algorithms needed to implement
RAID 4.
The benefits of RAID 4 disappear when you have many, small I/O operations
scattered randomly and widely across the disks in the array. RAID 4’s fixed parity disk becomes a bottleneck in such applications, as the following example illustrates. Let’s say the host instructs the controller to make two small writes. The
writes are widely scattered, involving two different stripes and different disk
drives. Ideally, you would like both writes to take place at the same time, but
RAID 4 makes this impossible, since the writes must take turns accessing the
fixed parity drive. For this reason, RAID 5 is the better choice for widely scattered, small write operations.
RAID 4 can withstand a s ingle failur e and handle
I/O acti vity without interrupt ion in degraded mode
until the failed drive is rebuilt. If a second drive
fails while the RAID s et is i n degraded mode, t he
entire RAID set will fail.
2.6.5 RAID 5
RAID 5 addresses the bottleneck issue for barrages of widely scattered, small I/O
operations. Like RAID 4, RAID 5 breaks up data into chunks, calculates parity,
and then writes the chunks in stripes to the disk drives, saving one drive one each
stripe for the parity data. Unlike RAID 4, however, RAID 5 changes the parity
drive on each stripe. This means, for instance, that a write operation involving
drive 2 on stripe 1 can conceivably take place at the same time as a write involving drive 3 on stripe 2, since they would be addressing different parity drives.
2–16 EK–SMCPO–UG. A01
CAUTION
Figure 2–8 Diagram of a RAID 5 Write
Chapter 2. RAID Array Controller
RAID 5 handles drive failures in the same manner as RAID 4, except the parity
is different for each stripe. The controller either uses the parity information on a
stripe to reconstruct its data or simply reads the data as normal, depending on the
location of the stripe’s parity drive.
While RAID 5 is ideally suited for applications with many, small I/O operations,
there is no reason why it cannot function equally well for applications with large,
sequential I/O operations. This makes RAID 5 an excellent all-purpose RAID
level.
EK–SMCPO–UG. A01 2–17
RAID Array 3000 Pedestal Enclosure
CAUTION
RAID 5 can withstand a s ingle failur e and handle
I/O acti vity without interrupt ion in degraded mode
until the failed drive is rebuilt. If a second drive
fails while the RAID s et is i n degraded mode, t he
entire RAID set will fail.
2.6.6 JBOD
JBOD, which stands for “Just a Bunch of Disks”, makes it possible to connect
one or standalone disk drives to the controller. A JBOD disk drive is not part of a
redundancy group, even though the controller assigns a redundancy group number to the drive. This number becomes that logical unit number (LUN) that the
host will use to address the drive.
One use for JBOD is to connect a system disk drive to the controller. The drive
does not become part of a RAID set, but it is made available to the host on the
same SCSI bus as the other devices controlled by the controller.
2.7 System Parameters
The system parameters (see Table 2–6) allow the user to customize certain aspects of the controller via the StorageWorks Command Console (SWCC)
Graphical User Interface.
2–18 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
g
g
y
g
g
g
g
g
g
g
g
g
y
g
g
g
y
Table 2–6 System Parameters
Parameter Description
Password Checking Enables or disables password checking. When password checkin
is enabled, the controller will limit access to certain options unless
the correct password is supplied.
Rebuild Rate Determines how much of the controller’s processing power is to be
used durin
rebuild operation, but ma
ber will not impact performance, but will require lon
the RAID set.
Create Rate Similar to Rebuild Rate, but determines the amount of processin
power to be used while creating a RAID set.
UPS Installed Determines whether the controller should enable the UPS warnin
signals (AC Lost and Two-Minute Warning). See “Backup Power
Management” for more details.
Read-Ahead Enabled Determines whether the controller should automatically pre-stage
data into cache durin
proves sequential access times as it improves the chance of
a cache hit on a read. However, it is not well suited for random
tin
access patterns since it relies on spatial locality.
Read-Ahead Limit Limits the amount of cache blocks that will be read ahead when
read ahead is enabled. This limit prevents the cache from bein
dominated by read-ahead data.
Validation Delay Determines how long the controller will delay after spinning up
drives before scannin
the drives have to initialize themselves.
12-Volt Sensor Enabled Determines whether the controller should monitor the 12-volt input
line.
Host Select WCE Specifies whether the controller allows the host to set and/or clear
the Write Cache Enable bit in the cachin
host s
formance. See “Write Operations” for more details.
Drive Command Time-out The number of seconds the controller will wait for a drive to re-
select after disconnectin
drive exceeds this time, the command will be retried. This value
also indicates the maximum amount of time a drive can be on the
SCSI bus. If a drive is on the bus for a
drive will be reset and the command will be retried.
Reset Propagation Specifies whether the host channels should propagate resets on
other host channels. This is useful in clusterin
where one cluster ma
cluster.
rebuild operations. A higher number speeds up the
impact I/O performance. A lower num-
er to rebuild
read operations. Enabling read ahead im-
them. The larger the number, the more time
mode page. Prevents
stems from turning write caching off and degrading per-
during the processing of a command. If a
reater length of time, the
environments
need to silence the adapter of a failing peer
et-
EK–SMCPO–UG. A01 2–19
RAID Array 3000 Pedestal Enclosure
2.8 Redundant Operation
When operating in a redundant configuration, the two controllers are linked such
that, in case of a failure, the surviving controller can access the other controller’s
cache memory and complete all operations that were in progress when the failure
occurred. The controllers support two different configurations:
•
ACTIVE / ACTIVE
: One host port is active on each controller. The other
port on each controller is passive and only used if the peer controller fails.
•
ACTIVE / PASSIVE
: Both host ports on one controller are active. The
other controller’s ports are both passive and only used if the primary controller fails.
When one controller fails, the survivor will process all I/O requests until the
failed controller is repaired and powered on. The subsystem will then return to its
previous state (i.e., ACTIVE / ACTIVE or ACTIVE / PASSIVE).
2.8.1 Initialization
During initialization, the firmware in the RAID 3000 verifies that both controllers have consistent configurations including identical memory cache and system
parameters. If the controller setups are incompatible, the set is not bound and
each controller operates in stand-alone mode.
2.8.2 Message Passing
Information is shared between the two controllers by a collection of messages
passed through the backplane connectors. The messages provide configuration
data as well as a heartbeat which is transmitted by each controller every 500 ms.
If a controller does not receive a heartbeat within one second, it assumes the peer
controller has become inoperable and begins failing over.
If the controllers cannot exchange messages due to communication problems
over the backplane, they will break the connection and each controller will
switch to a stand-alone mode.
2.8.3 Failover
Failover describes the process of transferring data from a failed controller to a
survivor and completing any active tasks. When one controller begins the failover process, it sends a reset to the other controller, which prevents the failing
unit from processing any more information and enables any host ports that are
passive. It then downloads the failed controller’s cache to its unused portion of
cache and begins acting upon that data.
2–20 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
While downloading the data, the controller responds to I/O by disconnecting (if
allowed) and waiting approximately three seconds before reconnecting and presenting a BUSY status. The delay is to prevent host operating systems from
seeing too many errors and fencing off the controller.
2.9 Environmental
The controller incorporates a set of on board sensors to detect abnormal operating conditions that may affect data safety.
2.9.1 Backup Power Management
The controller must be connected to a backup battery and/or Uninterruptable
power supply (UPS) to prevent the subsystem cache from being corrupted during
unexpected losses of power. If no backup power supply is provided, the controller will remain off-line and reject all I/O requests with a status of Check Condition/Hardware Error.
If the backup power source indicates that power may be failing, the controller’s
first step is to sound an alarm, enter write through mode and begin flushing
cache. If the backup power source reaches a critical state, the alarm frequency
increases and all host channels are disabled to prevent new requests from interfering with the cache flush.
Table 2–7 shows how the controller reacts to the power supplies.
Table 2–7 Backup Power Management
Battery Status
OFFLINE
(Not Installed)
NORMAL
WARNING
SEVERE
OFFLINE
(Not Installed)
OFFLINE NORMAL SEVERE OFFLINE
NORMAL NORMAL WARNING WARNING
SEVERE WARNING WARNING SEVERE
OFFLINE WARNING SEVERE OFFLINE
NORMAL
No Alarms
•
Normal I/O
•
EK–SMCPO–UG. A01 2–21
UPS Status
NORMAL WARNING
(AC Lost)
SEVERE
(2-Min. Warning)
RAID Array 3000 Pedestal Enclosure
WARNING
Slow Alarm
•
Normal I/O
•
SEVERE
UltraSCSI Alarm
•
Flushes Cache
•
Enters Write-Through Mode
•
OFFLINE
UltraSCSI Alarm
•
Flushes Cache
•
No New I/O Allowed
•
2.9.2 Voltage Monitoring
2.9.2.1 System Voltage
The controller monitors the incoming system voltage levels and ensures they are
satisfactory for controller operation. The acceptable voltage levels are listed in
Table 2–8.
Table 2–8 Acceptable System Voltage Levels
State Range Action
Normal 4.80 - 5.25 V Normal Operation
Low Warning 4.75 - 4.80 V Alarm
High Warning 5.25 - 5.30 V Alarm
Low Severe Less than 4.75 V Off-line
High Severe Greater than 5.30 V Off-line
2–22 EK–SMCPO–UG. A01
Chapter 2. RAID Array Controller
2.9.2.2 Termination Voltage
The controller monitors the incoming termination voltage levels and ensures they
are satisfactory for controller operation. The acceptable voltage levels are shown
in Table 2–9.
Table 2–9 Acceptable Termination Voltage Levels
State Range Action
Normal 4.20 - 5.40 V Normal Operation
Low Warning 4.00 - 4.20 V Alarm
High Warning 5.40 - 5.55 V Alarm
Low Severe Less than 4.00 V Off-line
High Severe Greater than 5.55 V Off-line
2.9.2.3 12 V Supply Voltage
The controller monitors the incoming 12-volt levels and ensures they are satisfactory for operation. The acceptable voltage levels are shown in Table 2–10.
Table 2–10 Acceptable 12 Volt Levels
State Range Action
Normal 10.80 - 13.80 V Normal Operation
Low Warning 10.20 - 10.80 V Alarm
High Warning 13.80 - 14.40 V Alarm
Low Severe Less than 10.20 V Off-line
High Severe Greater than 14.40 V Off-line
EK–SMCPO–UG. A01 2–23
RAID Array 3000 Pedestal Enclosure
2.9.3 Temperature Monitoring
2.9.3.1 External Temperature
The controller monitors the external operating temperature and ensures they are
satisfactory for controller operation. The acceptable temperature levels are
shown in the following table.
Table 2–11 Acceptable External Temperature Voltage Levels
State Temperature Action
Normal < 41º C Normal Operation
High Warning < 41º C - 52º C Alarm
High Severe > 52º C Off-line
2.9.3.2 Board Temperature
The controller monitors the on-board operating temperature and ensures they are
satisfactory for controller operation. The acceptable temperature levels are
shown in the following table.
Table 2–12 Acceptable Board Temperature Voltage Levels
State Temperature Action
Normal < 52º C Normal Operation
High Warning 52º C - 56º C Alarm
High Severe > 56º C Off-line
2–24 EK–SMCPO–UG. A01
3
Maintenance
This chapter describes how to interpret the status of the LEDs on the pedest al and use
them as a troubleshooting aid during a pedestal malfunction. Both the pedestal LEDs
and the LEDs on the major components are cover ed. The chapt er also describes how to
replace a Field Replaceable Unit (FRU) and how to reconfigure the SCSI bus.
3.1 Introduction
Troubleshooting the pedestal consists of monitoring the status of the external
LEDs to determine if a major component is malfunctioning. The Field Replaceable Units (FRUs) in the pedestal are:
Disk drives (SBBs)
•
RAID array controller
•
Power supplies
•
Environmental Monitor Unit (EMU) circuit board
•
Uninterruptable Power Supply (UPS)
•
The information in this chapter is divided into the following major sections:
Pedestal status and power supply LEDs
•
SBB status LEDs
•
Controller LEDs
•
EMU error reporting
•
Replacing an FRU
•
Reconfiguring the UltraSCSI bus
•
Replacing the controller memory cache modules
•
EK–SMCPO–UG. A01 3–1
RAID Array 3000 Pedestal Enclosure
3.2 Pedestal Status and power LEDs
The pedestal is equipped with two front panel LEDs (see Figure 3–1) that monitor the following error conditions:
A power supply fan that is not operating
•
An over-temperature condition
•
A dc power problem
•
External fault conditions
•
Controller faults
•
When the pedestal is operating properly, the green
ber pedestal
LED is off. Also, the green power supply LED on the rear of
fault
each power supply is on.
Figure 3–1 Pedestal Status LEDs
LED is on and the am-
power
3.3 SBB Status LEDs
Each disk drive SBB in the pedestal has two status LEDs – a green
and an amber
properly, the activity LED is flashing indicating normal disk activity on the SCSI
bus, and the
and recommends corrective actions when a LED indicates a fault condition.
3–2 EK–SMCPO–UG. A01
LED as shown in Figure 3–2. When the pedestal is operating
fault
LED is off. Table 3–1 summarizes the states of the SBB LEDs
fault
activity
LED
Chapter 3. Maintenance
Table 3–1 Disk Drive SBB Status LEDs
Activity LED Fault LED Indication
On Off Drive is operating properly.
Off Off Drive is inactive and operating normally. There is no
fault.
On On Fault status: drive is defective. Recommend that you
replace the device.
Off On Fault status: drive is inactive and not spinning. Rec-
ommend that you replace the device.
On Flashing Fault status: drive is active and slowing down due to
the fault.
Figure 3–2 Disk Drive Status LEDs
3.4 Controller LEDs
The LEDs on the front panel of the controller monitor host and disk channel activity and a controller fault condition. The reset button/LED flashes green approximately once every second (heartbeat) to indicate that the controller is operating normally. Figure 2–3, Chapter 2, identifies the LEDs. Table 2–2 describes their functions.
EK–SMCPO–UG. A01 3–3
RAID Array 3000 Pedestal Enclosure
3.5 EMU Error Reporting
The primary function of the EMU is to detect and report conditions that can
cause the pedestal to malfunction and to report malfunctions. To accomplish this
the EMU constantly monitors the following pedestal signals:
+5 and +12 V dc
•
ac input
•
Power Supply present
•
Total power
•
Power OK (P_OK)
•
Power Disabled
•
Fan Speed (minimum and high speed)
•
Fan exhaust temperature
•
Shelf OK (S_OK)
•
The EMU can exchange signals with auxiliary devices and controllers.
3.5.1 EMU Error Conditions
The EMU reports error conditions and malfunctions using an audible alarm and a
LED. The user-enabled audible alarm and the amber fault LED on the front of
the pedestal are the only error indicators.
Whenever any of the following error conditions occur, the amber fault LED on
the front of the pedestal will turn on. When alarm switch S1 on the rear panel of
the pedestal is in the up (enabled) position, the audible alarm will sound whenever one of the following conditions occurs:
Loss of ac power to one of the power supplies
•
Failure of either power supply fan
•
UPS not connected
•
UPS power failure
•
UPS output too low
•
Temperature exceeds 50º C (123º F)
•
One of the +12 V dc outputs is less than + 9.85 V dc
•
One of the +5 V dc outputs is less than + 3.95 V dc
•
Miscellaneous error condition
•
A controller error condition exists
•
External Fault
•
3–4 EK–SMCPO–UG. A01
Chapter 3. Maintenance
3.6 Replacing Components (FRU's)
This section describes how to replace an FRU in the RAID Array 3000 pedestal.
The information is divided into the following subsections:
Removing the pedestal door
•
Replacing an SBB
•
Replacing a power supply
•
Replacing the controller
•
Replacing the EMU board
•
Replacing the UPS
•
Replacing a controller memory cache module
•
3.6.1 Removing the Pedestal Door
Proceed as follows to remove the pedestal door (Figure 3–3):
1. Unlock and open the door to a 90º angle in relation to the pedestal.
2. Carefully lift up on the door until the hinge pins are against the top of
the mounting holes.
3. Pull the door straight out until the hinge pins clear the bezel.
Figure 3–3 Removing Pedestal Door
EK–SMCPO–UG. A01 3–5
RAID Array 3000 Pedestal Enclosure
3.6.2 Replacing an SBB
There are two methods for replacing a disk drive SBB with an
•
Hot swap
and installing SBBs while the bus is active. Hot swap is supported by the
RAID Array 3000 controller
•
Cold swap
UPS to deactivate the bus
Perform the following procedure to replace an SBB in the pedestal
(see Figure 3–4):
– This method requires that the SCSI controller support removing
– Requires removing ac power from the pedestal and disabling the
CAUTION
• Always use both hands when remov ing or i nserting an SBB.
• Never remove a drive fr om the pedestal before it has completely spun down.
• Take care not to s lam the replacement SBB
into the pedestal enclosure.
identical
SBB:
1. Unlock and open the front door.
2. Release the drive from the pedestal slot by squeezing the mounting tabs on
the SBB together, but do not remove the SBB from the pedestal.
3. Wait 15-30 seconds for the drive to stop spinning before removing it from
the pedestal slot.
4. Insert an identical model SBB in the slot and push it in until an audible click
is heard indicating the SBB is fully seated (SBB front panel mounting tabs
expand and engage the pedestal shelf).
5. Observe that the SBB status LEDs are operating.
6. Close the front door.
3–6 EK–SMCPO–UG. A01
Figure 3–4 Replacing an SBB
Chapter 3. Maintenance
3.6.3 Replacing a Power Supply
You can replace a power supply without affecting pedestal operation using the
following procedure:
CAUTION
When you remove a power supply, the airflow
through the SBBs is int er rupt ed. Al way s i ns tal l t he
replacement power supply as quickly as possi ble
to prevent overheating.
1. Grasp the power supply handle shown in Figure 3–5.
2. Slide and hold the locking latch to the left and pull the supply out using a
short jerking motion.
3. Insert the replacement power supply into the pedestal and carefully align it.
4. Push the power supply in until it is fully seated and the locking tab engages.
5. Until the fan is operating at the proper speed, the green power supply LED
remains off. When the fan reaches the proper speed (several seconds), the
LED should come on.
EK–SMCPO–UG. A01 3–7
RAID Array 3000 Pedestal Enclosure
Figure 3–5 Replacing a Power Supply
3–8 EK–SMCPO–UG. A01
Chapter 3. Maintenance
3.6.4 Replacing the RAID Array Controller
Perform the following procedure to replace the RAID Array controller (see Figure 3–6):
1. Grip the two locking latches on the front panel of the controller and pull
them forward until the controller disengages from its mounting slot.
2. Remove the controller by sliding it forward and free of the pedestal.
3. Insert the replacement controller into the open slot, align the module into the
card guides, and gently slide it into the pedestal until the connector engages
the backplane connector in the pedestal.
4. Turn the two front panel latches inward to fully seat the controller in the
pedestal slot.
Figure 3–6 Removing the Controller from the Pedestal
EK–SMCPO–UG. A01 3–9
RAID Array 3000 Pedestal Enclosure
3.6.5 Replacing the EMU Board
WARNING
Only qualified service personnel should replace
the EMU
the pedestal si de panel i s rem oved. Alway s power
off the pedest al and remove the power cords before replacing the EMU.
Perform the following procedure to replace the EMU Board:
Tools required:
Flat-blade or 3/16 (5 mm) hex-head screwdriver
•
# 10 TORX-head screwdriver
•
1. Quiesce the host bus by shutting down the host system.
To allow the UPS to supply power while the cache
is being flushed to disk, do not unplug the base
pedestal from the UPS.
2. Unplug the base pedestal power cord from the wall outlet.
3. Unplug the UPS power cord from the wall outlet, the UPS will now signal
the controller to flush the cache.
4. Wait until the UPS shuts down completely (this may take several minutes).
5. Power off the base pedestal and plug the UPS power cord into the wall outlet.
6. Remove the pedestal door.
7. Remove the left side cover from the pedestal (see section 2.3, Chapter 2).
8. Record the orientation of the color traces on the two ribbon cables connected
at the rear of the EMU board and disconnect the cables.
9. Remove the mounting nuts from the UPS and external fault connectors on
the rear panel using a hex-head or a flat-blade screwdriver (see Figure 3–7).
Remove the board from the pedestal.
. Dangerous v oltages are ex posed when
CAUTION
3–10 EK–SMCPO–UG. A01
Chapter 3. Maintenance
Figure 3–7 Remove Nuts from UPS and External Fault Connectors
3000-49
10. Align the connectors on the replacement EMU with the connector holes on
the rear panel and replace the mounting studs to secure the board (
over-tighten
).
11. Reconnect the two ribbon cables to the rear connectors on the EMU board.
12. Replace the side cover on the pedestal and secure it with the TORX-head
screw on the front bezel of the pedestal.
13. Replace the door on the pedestal.
EK–SMCPO–UG. A01 3–11
do not
RAID Array 3000 Pedestal Enclosure
3.6.14 Replacing the UPS
Proceed as follows to replace the UPS:
1. Ensure the UPS power switch is set to off.
2. Disconnect the pedestal power cord from the UPS.
3. Disconnect the UPS power cord from the wall outlet.
4. Disconnect the signal control cable from the UPS
5. Install the replacement UPS and reconnect the power cords and the signal
control cable.
6. Set the UPS power switch to on and then power up the pedestal.
3.7 Differential/Wide UltraSCSI Bus
The reliability of data transfers on a SCSI bus depends on the following factors:
The bus transmission rate
•
The maximum SCSI bus length (a function of bus type, the transmission rate,
•
and the use of SCSI bus converters)
The total length of the SCSI bus (as measured from the host bus terminator to
•
the subsystem terminator)
MT/s (mega-transfers per second) is the repetitive rate at which words of data
•
are transferred across a bus. The number of megabytes per second (MB/s) is
determined by the bus width (8 or 16-bit) and the number of bytes per word
(1 or 2, respectively)
Table 3–2 lists the maximum SCSI bus lengths and the longest SCSI
cables recommended by DIGITAL.
Because the bus length includes the cable plus
the backplane distance inside the enclosure, the
recommended max imum cable lengths listed
not
the same as the maximum bus length.
3–12 EK–SMCPO–UG. A01
NOTE
are
Chapter 3. Maintenance
Table 3–2 SCSI Bus Length and External Cables
Bus Rate Bus Length Longest DIGITAL Cable
Speed MT/s MB/s Meters Feet Number Meters Feet
Fast 20 40 25 82 BN21K-23
BN21L-23
23 82
The SCSI bus in your pedestal is factory-configured as a split bus. One bus is
designated bus D0 and the other as bus D1. An internal eight-position step switch
in the cabinet controls the SCSI bus device address configuration of the SBB
slots in the pedestal. For the RAID Array 3000, the switch is preset at the factory
to configuration “6”. This results in a device slot address assignment of 8 through
11 for bus D0 and 8 through 10 for bus D1. Table 3–3 shows a listing of the device addresses for each bus and their corresponding pedestal slot location.
Table 3–3 Assigned Slot Device Addresses in the Pedestal
Slot #
0 12 3 456
Bus D0 D1
Device
8 9 10 11 8 9 10
Address
3.7.1 Reconfiguring the SCSI Bus
WARNING
Only quali fied servi ce personnel s hould reconfigure the SCSI bus
within the subsystem. To prevent electrical shock,
always turn the subsystem off and disconnect the
power cords before removing the side panel.
If you want to reconfigure the SCSI bus (to add the Expansion Pedestal Option to
your subsystem installation for example) you must reconfigure the bus. This involves powering down the subsystem installation, removing the left side panel,
and reconfiguring the bus by changing the setting of the configuration switch.
Adding the Expansion Pedestal Option to your installation is described in Chapter 4 of this guide.
. Dangerous voltages ar e present
EK–SMCPO–UG. A01 3–13
RAID Array 3000 Pedestal Enclosure
The SCSI bus configuration switch selects the eight (0 – 7) SCSI bus configurations in the pedestal. Each bus configuration determines the slot device addresses
(0 – 6) for both 8-bit and 16-bit devices. The pedestal is configured at the factory
for split-bus operation (configuration switch set to “6”). To change the configuration of the bus, you must remove the left side panel to gain access to the backplane and the configuration switch.
NOTE
SCSI device addresses 6 and 7 are reserved for
the RAID Array controllers. The top controller slot
in the pedestal is assigned device address 7 and
the bottom slot (redundant controller) is device address 6. The controller allows a maximum of 15 ( or
14 with dual controllers) disk devices on each of its
device buses.
To remove the side panel, proceed as follows:
1. Quiesce the host bus by shutting down the host system.
CAUTION
To allow the UPS to supply power while the cache
is being flushed to disk, do not unplug the base
pedestal from the UPS.
2. Unplug the base pedestal power cord from the wall outlet.
3. Unplug the UPS power cord from the wall outlet, the UPS will now signal
the controller to flush the cache.
4. Wait until the UPS shuts down completely (this may take several minutes).
5. Power off the base pedestal and plug the UPS power cord into the wall outlet.
6. Disconnect the SCSI cable from the host connector on the rear panel.
7. Remove the front door on the pedestal.
8. Remove the cover screw from the left side panel (see Figure 3–8).
9. Grasp the handhold at the rear of the panel and pull the panel toward the rear
and clear of the cabinet.
3–14 EK–SMCPO–UG. A01
Chapter 3. Maintenance
NOTE
To decrease the configuration number:
the upper switch to st ep the address down one address at a time (decrement) until the desired configuration number is displayed.
To increase the configuration number:
lower s w i t c h to step t h e add ress up on e ad d r ess at
a time (increment) until the desired configuration
number is displayed.
Press
Press the
9. Figure 3–9 shows the location of the switch. Figure 3–10 depicts the separate
step switches to set the subsystem to the desired SCSI bus configuration
(refer to the label at bottom of pedestal to cross reference slot SCSI ID addresses with switch settings).
Figure 3–8 Remove Screw and Panel
EK–SMCPO–UG. A01 3–15
RAID Array 3000 Pedestal Enclosure
300
0-3
8
Figure 3–9 Location of SCSI Bus Configuration Switch
6
Figure 3–10 Configuration Switch
6
NOTE
Reconfiguring the bus also involves cable and bus
terminator changes. Chapter 4 of this guide describes in detail how to reconfigure the bus.
3–16 EK–SMCPO–UG. A01
Chapter 3. Maintenance
3.8 Replacing the Controller Memory Cache Modules
The two memory cache modules in the RAID controller are replaced by removing the controller from the pedestal to gain access to the modules. Then, place
the controller on a flat working surface and proceed as follows:
CAUTION
To prevent an electr ical discharge from damaging
the S IMM s, al ways wear an E SD wris t str ap c onnected to a suitable ground when handling the
memory chips.
1. Power down the pedestal first and then the UPS.
2. Grasp the latches on the front of the controller and pull them forward until the controller disengages from the pedestal (see Figure 3–11).
3. Remove the controller from the pedestal and place on a flat working surface.
4. Remove the two installed cache memory SIMM modules by releasing the locking
clips at each end of the module until it disengages and snaps into an upright position as shown in Figures 3–12 and 3–13.
Figure 3–11 Remove Controller
EK–SMCPO–UG. A01 3–17
RAID Array 3000 Pedestal Enclosure
Figure 3–12 Release Locking Clips
Figure 3–13 Remove Installed SIMM Modules
Ensure the “side 1” side of the two replacement
SIMMs is facing toward you when installing the
modules in the following step.
3–18 EK–SMCPO–UG. A01
CAUTION
Chapter 3. Maintenance
5. Installed the two replacement memory modules by aligning the module and connector pins (check alignment guide in center of module) and gently pivot the module the main controller board until it snaps into place (see Figures 3–14 and 3–15).
6. Replace the controller into the pedestal.
7. Power on the UPS and pedestal and check the activity LEDs on the front panel of
the controller. The reset switch/LED should begin to flash at a half-second rate
(heartbeat) and the host activity LEDs should flash.
Figure 3–14 Install Replacement Modules
EK–SMCPO–UG. A01 3–19
RAID Array 3000 Pedestal Enclosure
Figure 3–15 Pivot Module Down to Secure
3–20 EK–SMCPO–UG. A01
4
Expansion Pedestal Option
This chapter describes the major feat ures of the Expansion Pedestal Opt ion and how to
connect the option to your RAID Array 3000-storage subsystem.
4.1 Product Description
The StorageWorks pedestal expansion option (Figure 4–1) is designed to expand
the storage capacity of the RAID Array 3000 subsystem. When connected to the
RAID Array 3000 base pedestal, the expansion option contains the basic components required to create a dual-pedestal storage array with a 16-bit, single-ended
Ultra-SCSI bus. The option enables a user to add up to seven 3½-inch SBBs to
create a 14-device storage array. The expanded array is controlled and operated
in an identical fashion as the base RAID Array 3000 subsystem.
Figure 4–1 Expansion Pedestal
EK–SMCPO–UG. A01 4–1
RAID Array 3000 Expansion Pedestal
4.2 Expansion Pedestal Cabinet
The expansion pedestal cabinet is a modular free-standing storage cabinet that is
completely self contained with dual fan-cooled power supplies, an internal UltraSCSI single-ended extender module, and an internal EMU circuit board. The
cabinet dimensions are the same as the subsystem base pedestal which houses the
controller and is normally installed within one meter of the base cabinet to facilitate the cable connections between the two units. Figure 4–2 identifies the expanded subsystem's SBB slots and they’re corresponding SCSI ID addresses.
Figure 4–3 identifies the items on the rear panel power supplies. The characteristics of the expansion pedestal cabinet are outlined below.
The storage device capacity of the expansion pedestal is seven 3½-inch SBBs
•
The pedestal slots are numbered 0 through 6 from top to bottom
•
There are seven SCSI bus device addresses (target IDs) 8 through 14 which
•
can be assigned to the 3½-inch SBBs
There is a single 68-pin, VHDC female SCSI connector on the rear panel
•
which interconnects the SBB expansion pedestal to the controller pedestal
The rear panel contains an alarm switch and an external fault condition con-
•
nector
The expansion pedestal contains two interchangeable fan-cooled ac power
•
supplies for redundant power
The expansion pedestal is equipped with an internal configuration switch
•
which selects one of the eight (0 through 7) subsystem configurations (set to
the correct configuration setting at the factory to properly integrate the expansion pedestal to the controller pedestal)
4–2 EK–SMCPO–UG. A01
Chapter 4. Expansion Pedestal Option
Figure 4–2 Expansion Pedestal Slot Locations and ID Addresses
SLOT 0
SLOT 1
SLOT 2
SLOT 3
SLOT 4
SLOT 5
SLOT 6
EXPANSION PEDESTAL
3000-19A
ID = 8
ID = 14
EK–SMCPO–UG. A01 4–3
RAID Array 3000 Expansion Pedestal
Figure 4–3 Rear Panel Power Supplies
4.3 Expansion Pedestal Components
The expansion pedestal contains a 16-bit, wide/differential UltraSCSI bus, an
Environmental Monitor Unit (EMU), a SCSI bus extender module, and two universal 50/60 Hz, 100 – 240 Vac fan-cooled power supplies.
The single-ended UltraSCSI bus is factory-configured as one continuous bus that
runs along the backplane between the disk drive connectors and the internal cables. These cables connect the drives to the connectors located on the rear panel
of the expansion pedestal. The device addresses on the bus are set at the factory
by an internal configuration switch. When set to a specific position, the switch
controls the addresses of each SBB slot.
The SCSI bus extender module extends the allowable electrical length of the bus
to accommodate longer physical SCSI cable connections between the base and
expansion pedestals.
4–4 EK–SMCPO–UG. A01
Chapter 4. Expansion Pedestal Option
300
0-3
The EMU (Figure 4–4) is an internal circuit board, which monitors the operation
of the pedestal. The EMU monitors power supply voltages, fans, temperatures
that are reported to the user, and controls (turns on and off) the audible alarm and
status LED on the front panel of the pedestal. It is connected to the SCSI bus and
powered by internal cabling. The following external components on the rear
panel of the expansion pedestal are part of the EMU board:
The alarm switch (S1) that enables (up) or disables (down) the audible alarm
•
The External Fault Condition connector allows the EMU to monitor the status
•
of a user-selected device
Figure 4–4 EMU Circuit Board Location
EM U
4.4 Reconfiguring Base Pedestal UltraSCSI Bus
Only quali fied servi ce personnel s hould reconfigure the base pedestal
pr esen t wit hi n t he s ub s ys t em. T o pr even t elec t ri cal
shock, always turn t he subsystem off and disconnect the power cords before removing the side
panel.
EK–SMCPO–UG. A01 4–5
7
WARNING
. Dangerous voltages are
RAID Array 3000 Expansion Pedestal
The RAID Array 3000 base pedestal is factory-configured for split-bus operation.
You must reconfigure the bus in the base subsystem from split-bus to a “throughbus” configuration prior to connecting the expansion cabinet to the base subsystem. The components needed to reconfigure the split-bus in the base subsystem
are included with your pedestal expansion kit option.
CAUTION
To prevent electrostatic discharge (ESD) from
damaging the controller, always wear an ESD wrist
strap connected to a suitable ground whenever you
handle any of the electronic components.
Perform the following procedure to reconfigure the SCSI bus in the base pedestal
from a “split-bus” to a “through-bus” configuration:
1. Perform an inventory of the bus conversion items supplied with the pedestal
expansion kit option. The items should consist of:
SCSI bus jumper cable 17-04166-03
•
SCSI cable 17-04454-01
•
2. Quiesce the host bus by shutting down the host system.
CAUTION
To allow the UPS to supply power while the cache
is being flushed to disk, do not unplug the base
pedestal from the UPS.
3. Unplug the base power cord from the wall outlet.
4. Unplug the UPS power cord from the wall outlet, the UPS will now signal
the controller to flush the cache.
5. Wait until the UPS shuts down completely (this may take several minutes).
6. Power off the base pedestal and plug the UPS into the wall outlet.
7. Remove the side cover from the base pedestal (Figure 4–5).
4–6 EK–SMCPO–UG. A01
Chapter 4. Expansion Pedestal Option
300
0-40
Figure 4–5 Remove Side Cover from Base Pedestal
8. Remove the bus terminator from backplane connector J11 (Figure 4–6).
Figure 4–6 Remove SCSI Bus Terminator
Remove Terminator
9. Remove the SCSI cable from device # 1 connector and backplane connector
J16 (see Figure 4–7).
EK–SMCPO–UG. A01 4–7
RAID Array 3000 Expansion Pedestal
300
0
Figure 4–7 Disconnect SCSI Cable
Remove SCSI Cable
-41
10. Connect jumper cable 17-04166-03 between the backplane connector J11
and the backplane connector J16 (see Figure 4–8).
Figure 4–8 Connect SCSI Jumper
J11
J16
Connect Jumper Cable
-
4–8 EK–SMCPO–UG. A01
Chapter 4. Expansion Pedestal Option
11. Remove the “knock-out” plate located above the D1 OUT label on the rear
panel of the base pedestal (see Figure 4–9).
Figure 4–9 Remove Connector Knockout Plate
Re move Kno c ko ut
from D 1 O u t
12. Connect cable assembly 17-04454-01 between the D1 OUT bulkhead opening and the device # 1 backplane connector (see Figure 4–10). Secure the
bulkhead connector by tightening the two 6-32 SEM screws.
EK–SMCPO–UG. A01 4–9
3000-45
RAID Array 3000 Expansion Pedestal
300
0-43
300
0
Figure 4–10 Connect SCSI Cable
Connect SCSI Cable
?
13. Set the bus configuration switch to “7” (see Figures 4–11 and 4–12). Figure
4–13 shows the reconfigured SCSI bus addresses of the expanded subsystem.
Figure 4–11 Set Configuration Switch to 7
4–10 EK–SMCPO–UG. A01
7
-44
Chapter 4. Expansion Pedestal Option
Figure 4–12 Configuration Switch
Figure 4–13 Reconfigured SCSI Bus Addresses
EK–SMCPO–UG. A01 4–11
RAID Array 3000 Expansion Pedestal
300
0
14. Reinstall the side panel on the base pedestal.
15. Connect the HD68-to-HD68 SCSI cable from the D1 OUT connector on the
base pedestal to the D1 IN connector on the rear of the expansion pedestal as
shown in Figure 4–14.
Figure 4–14 Connect SCSI Cable Between Pedestals
Base
Pedestal
Expans ion
Pedestal
16. Make the power cable connections between the expansion pedestal, the UPS, and
the ac power source (see Figure 4–15).
4–12 EK–SMCPO–UG. A01
-47
Figure 4–15 Power Cable Connections
300
0-5
0
Chapter 4. Expansion Pedestal Option
Connector
AC
Power
AC
Power
UPS
AC
Power
Base
Pedestal
Expa nsio n
Pedestal
UPS
If you want t o maintain t he exist ing R AID level c onfiguration, three drives must be relocated from the
base pedestal to specific slots in the expansion
pedestal as described in step 13.
17. Transfer the bottom three disk drives (slots 4, 5, and 6) from the base pedestal to
the top three slot locations (slots 0, 1, and 2) in the expansion pedestal as shown in
Figure 4–16.
EK–SMCPO–UG. A01 4–13
CAUTION
RAID Array 3000 Expansion Pedestal
Figure 4–16 Transfer Drives from Base to Expansion Pedestal
ID = 8
ID = 9
ID = 1 0
ID = 11
ID = 8
ID = 9
ID = 1 0
Base Pedestal
ID = 8
ID = 9
ID = 10
ID = 11
ID = 12
ID = 13
ID = 14
Expansion Pedestal
3000-46
18. Install the new drives in the remaining slots each pedestal to complete the
installation.
19. Power up the UPS and the two pedestals and then proceed to StorageWorks
RAID 3000 Configuration and Maintenance Guide EK–SMCS2–UG to configure the expanded subsystem.
4–14 EK–SMCPO–UG. A01
5
Second Controller Option
This chapter describes how to install a second RAID controller in the pedestal. The second controller provides a fail/safe feature to protect your data in case of a pr imary controller malfunction. The r edundant controller is installed directly below the primary (top)
controller.
5.1 Introduction
The controller option provides a second (redundant) controller unit in your subsystem installation to preserve the integrity of data should the first controller
sustain a malfunction. The second controller is installed directly below the existing device in the bottom controller slot of the pedestal (see Figure 5–1).
The installation procedure consists of adding two SIMM memory modules to the
redundant controller and one memory module to the existing controller in the
pedestal. Following the memory upgrade, you simply insert and seat the devices
in their respective controller slots and configure the subsystem to accommodate a
redundant controller. The following steps summarize the option installation procedure:
Ensure the SCSI bus is quiescent prior to powering off the host, the pedestal,
•
and the UPS
Install two of the SIMM modules supplied with the kit into the second con-
•
troller (all SIMMs must have the same memory capacity)
Install the third SIMM module supplied with the kit into the existing control-
•
ler
Install the controllers into their appropriate slots in the pedestal
•
Power up the subsystem and configure the second controller for redundant
•
controller operation
EK–SMCPO–UG. A01 5–1
RAID Array 3000 Pedestal Enclosure
300
0-48
Figure 5–1 Second Controller Slot Location
Second
Controller
Slot
5.2 Installation Procedure
To prevent an electr ical discharge from damaging
the S IMM s, al ways wear an E SD wris t str ap c onnected to a suitable ground when handling the
memory modules.
1. Ensure that the UltraSCSI bus is in a quiescent state (no I/O activity).
2. Shut down the host system.
3. Issue a “shutdown” command from the SWCC console to the pedestal controller.
4. Power off the pedestal(s) and the UPS.
5–2 EK–SMCPO–UG. A01
CAUTION
Chapter 5. Second Controller Option
5. Perform an inventory of the items in the second controller kit. The kit should
contain the following:
RAID Array 3000 controller
•
memory SIMMs
•
Model label
•
Warranty Card
•
6. Install two of the SIMM modules into the second controller (make sure all
modules are of the same type) by aligning the connector pins and inserting
the modules into the SIMM module connectors as shown in Figure 5–2.
Figure 5–2 Insert Module into SIMM Connector
7. Ensure the module is firmly seated and then gently pivot it toward the controller board until it snaps into place as shown in Figure 5–3.
EK–SMCPO–UG. A01 5–3
RAID Array 3000 Pedestal Enclosure
Figure 5–3 Pivot Module Down to Seat
8. Remove the existing controller (see Figure 5–4) and install the third module
into the empty SIMM connector using the same procedure described in step
6.
Figure 5–4 Remove Controller from Top Slot
5–4 EK–SMCPO–UG. A01
Chapter 5. Second Controller Option
9. Replace the existing controller into the top controller slot in the pedestal and
seat it into place by pushing forward on the side quick-disconnect handles.
Make sure the guides on each side of the controller align with the guides in
the slot.
10. Remove the dummy filler panel from the bottom controller slot by pulling
the handle straight out.
11. Install the second controller into the bottom controller slot and seat it into
place by pushing forward on the side quick-disconnect handles.
12. Power up the subsystem and refer to the RAID Array 3000 configuration and
maintenance guide (EK–SMCS2–UG) for information describing how to
configure the redundant controller option.
EK–SMCPO–UG. A01 5–5
Reader’s Comments
Manual Order Number:
EK–SMCPO–UG. A01
RAID Array 3000 Storage Subsystem Hardware User's Guide
Digital is committed to providing the best possible products and services. Since our manuals are
important components of our products, we value your comments, corrections, and suggestions
for improvements. Please take a few minutes to fill out and return this form, attaching additional
sheets, if needed. Thank you.
Manual Rating Excellent Good Fair Poor
Accuracy (correct presentation of facts) [ ] [ ] [ ] [ ]
Completeness (adequate information) [ ] [ ] [ ] [ ]
Clarity (easy to understand) [ ] [ ] [ ] [ ]
Organization (logical sequence of information) [ ] [ ] [ ] [ ]
Layout (easy to follow subject matter) [ ] [ ] [ ] [ ]
Indexing (easy to locate desired information) [ ] [ ] [ ] [ ]
Errors Noted
Return Address:
Customer Research
Response Center
Digital Equipment Corporation
334 South Street, SHR3-2/S27
Shrewsbury, MA 01545
(please include page, paragraph, table or figure number)
Name Phone
Title
Company
Street Address
Mail Stop
City State ZIP
Country (if other than USA)
Loading...